NVIDIA and Microsoft Tease 'New Era of PC' at Computex — N1X ARM Chip Confirmed for June 1
NVIDIA and Microsoft posted coordinated 'A new era of PC' teasers on May 29, pointing to Jensen Huang's GTC Taipei keynote at Computex where the co-developed N1X Windows-on-Arm SoC is expected to officially launch.
NVIDIA and Microsoft published identical “A new era of PC” posts on X on May 29, both tagged with GPS coordinates pointing to Taipei Music Center — the venue for Jensen Huang’s keynote at GTC Taipei during Computex 2026, scheduled for June 1. The coordinated move is the clearest signal yet that the NVIDIA/MediaTek N1X Windows-on-Arm SoC is ready to launch.
According to leaked benchmarks and supply chain reporting, the N1X pairs a 20-core Arm CPU with a Blackwell-class GPU on TSMC 3nm, targeting ultrathin laptops in the $1,000–$1,800 range. For NVIDIA, this is its first attempt to own the CPU-plus-GPU layer on ARM Windows hardware — historically, its territory has been discrete GPUs and data center accelerators. If the chip delivers on leaked tokens-per-watt claims, it would reshape the Windows laptop market entirely.
Three OEMs are confirmed to launch N1X devices this holiday cycle: Asus (Vivobook N1X 15), Lenovo (Slim 7x AI), and Dell (XPS 13 Aero). All three appeared in FCC and Eurasian Economic Commission filings over the past six weeks.
Since Apple released the M1 in 2020, no Windows PC chip has genuinely threatened Qualcomm’s Snapdragon X Elite in ARM silicon. The N1X is the first challenger with the GPU pedigree to match Apple’s integrated approach. NVIDIA’s driver ecosystem and CUDA tooling are already on Windows; plugging in a Blackwell GPU die means developers can target the NPU via TensorRT-compatible paths, not Qualcomm’s Hexagon stack.
Jensen Huang is expected to keynote on June 1 alongside Satya Nadella joining remotely. Microsoft’s Windows account has not posted coordinated hardware teasers since the 2022 Surface event — doing so now signals a formal co-announcement, not a partner endorsement.
For developers, the N1X means NPU fragmentation on Windows is accelerating. The Snapdragon X Elite uses Hexagon NPU instructions; the N1X uses NVIDIA’s TensorRT-compatible path. Apps that call Windows ML or ONNX Runtime directly will run without modification. Custom CUDA or Tensor Core code needs porting from Qualcomm’s stack. This will matter as more inference-at-the-edge workloads target Copilot+ certified hardware.
Computex 2026 runs May 30 through June 6. Full spec sheets, device pricing, and driver availability are expected June 1.